Ninety percent of stroke survivors have clinically significant gait impairments that lead to secondary medical complications, including cardiovascular deconditioning and reduced quality of life (QOL). Several rehabilitative interventions that increase the level of activity and mobility have been shown to be beneficial. Challenges posed by most of these locomotor gait training interventions require recurring visits to a rehabilitation center. The recent development of novel powered exoskeletons offers a potential mechanism for stroke survivors to improve mobility in the home and community. Although the predominant research using these devices have been in persons with spinal cord injury, there are currently two devices approved by the FDA for use in patients with stroke and additional devices are being developed. The Keeogo powered orthotic exoskeleton is a novel device intended for persons with stroke who can ambulate but have gait impairment. This device consists of a ridged orthotic structure placed over clothing on the legs and batteries to the power motors that assist both knees in gait movement. The system monitors hip movement driven by the user and interprets this movement to apply the appropriate assistance at the knee joint. This unique approach makes learning intuitive, enabling the user only to acclimate to the system rather than learning how to control the device to initiate the desired movement. This proposal is a randomized controlled pilot study. Fifteen veterans with chronic stroke (>6months) and who retain some ability to take steps but have impaired gait will be recruited. Ten participants will be randomized into the exoskeleton group and 5 into the control group. Both groups will be asked to complete 36 one-hour sessions of ambulation training. The exoskeleton group will train using the Keeogo powered orthotic exoskeleton and the control group will train without using their own conventional aide. The primary aim is to determine the efficacy of training with this exoskeleton and its ability to improve transfers to standing and sitting as assessed by the five times sit-to-stand test. A secondary aim will be to assess effects of the device during overground ambulation. Outcome measurements to investigate changes of knee range of motion and loading of the paretic limb will be accomplished using an instrumented goniometer; changes in limb loading will be monitored using a foot pressure mapping system; and stepping parameters will be recorded using the GaitRite carpet. An exploratory outcome of changes in energy expenditure during ambulation with and without the powered exoskeleton will be assessed from a 6-minute walking test using a portable metabolic cart. Additional exploratory outcome measures of QOL will be determined using the SF-36 and the Stroke Specific Quality of Life Scale. Baseline testing for both groups will be performed while ambulating without the device and prior to starting in the training program. The exoskeleton group will repeat testing within the first 3 sessions of the training program. Post assessments will be obtained after 36 sessions of training. Retention will be assessed by repeating the assessments without the exoskeleton after 1 month of stopping the training program. It is anticipated that longitudinal training with the device will increase number of squats achievable, increase knee flexion during swing, increase loading of the paretic limb, and induce symmetrical stepping. No immediate effect on energy expenditure by the device is expected however, a training effect is expected to demonstrate improvement in the energy cost of ambulation. Any gains observed during this study may not be retained, but as this device has the potential to be used in the home and community environment, justifying a prescription of this device for retention purposes may be warranted. Use of this device may also lead to improvement of QOL.